Macs2 calllpeak (#1038)

* Add tests and yml file for macs2/callpeak

* add format option for macs2

* update macs2/callpeak to accept format argument

* update test.yml

* update the container version.

* try to fix the issue in conda container.

* Update conda and containers

* Going back to previous container versions

Co-authored-by: JoseEspinosa <kadomu@gmail.com>

modules/macs2/callpeak/main.nf

@@ -13,9 +13,9 @@ process MACS2_CALLPEAK {
 
     conda (params.enable_conda ? "bioconda::macs2=2.2.7.1" : null)
     if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
-        container "https://depot.galaxyproject.org/singularity/macs2:2.2.7.1--py38h0213d0e_1"
+        container "https://depot.galaxyproject.org/singularity/macs2:2.2.7.1--py38h4a8c8d9_3"
     } else {
-        container "quay.io/biocontainers/macs2:2.2.7.1--py38h0213d0e_1"
+        container "quay.io/biocontainers/macs2:2.2.7.1--py38h4a8c8d9_3"
     }
 
     input:
@@ -33,12 +33,19 @@ process MACS2_CALLPEAK {
 
     script:
     def prefix   = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+    def args     = options.args.tokenize()
     def format   = meta.single_end ? 'BAM' : 'BAMPE'
     def control  = controlbam ? "--control $controlbam" : ''
+    if(args.contains('--format')){
+        def id = args.findIndexOf{it=='--format'}
+        format = args[id+1]
+        args.remove(id+1)
+        args.remove(id)
+    }
     """
     macs2 \\
         callpeak \\
-        $options.args \\
+        ${args.join(' ')} \\
         --gsize $macs2_gsize \\
         --format $format \\
         --name $prefix \\

modules/macs2/callpeak/meta.yml

@@ -0,0 +1,63 @@
+name: macs2_callpeak
+description: Peak calling of enriched genomic regions of ChIP-seq and ATAC-seq experiments
+keywords:
+    - alignment
+    - atac-seq
+    - chip-seq
+    - peak-calling
+tools:
+  - macs2:
+      description: Model Based Analysis for ChIP-Seq data
+      homepage: None
+      documentation: https://docs.csc.fi/apps/macs2/
+      tool_dev_url: https://github.com/macs3-project/MACS
+      doi: "https://doi.org/10.1101/496521"
+      licence: ['BSD']
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - ipbam:
+      type: file
+      description: The ChIP-seq treatment file
+  - controlbam:
+      type: file
+      description: The control file
+  - macs2_gsize:
+      type: string
+      description: Effective genome size. It can be 1.0e+9 or 1000000000, or shortcuts:'hs' for human (2.7e9),
+        'mm' for mouse (1.87e9), 'ce' for C. elegans (9e7) and 'dm' for fruitfly (1.2e8)
+
+output:
+  - versions:
+      type: file
+      description: File containing software version
+      pattern: "versions.yml"
+  - peak:
+      type: file
+      description: BED file containing annotated peaks
+      pattern: "*.gappedPeak,*.narrowPeak}"
+  - xls:
+      type: file
+      description: xls file containing annotated peaks
+      pattern: "*.xls"
+  - gapped:
+      type: file
+      description: Optional BED file containing gapped peak
+      pattern: "*.gappedPeak"
+  - bed:
+      type: file
+      description: Optional BED file containing peak summits locations for every peak
+      pattern: "*.bed"
+  - bdg:
+      type: file
+      description: Optional bedGraph files for input and treatment input samples
+      pattern: "*.bdg"
+
+authors:
+  - "@ntoda03"
+  - "@JoseEspinosa"
+  - "@jianhong"

tests/config/pytest_modules.yml

@@ -778,6 +778,10 @@ lofreq/indelqual:
   - modules/lofreq/indelqual/**
   - tests/modules/lofreq/indelqual/**
 
+macs2/callpeak:
+  - modules/macs2/callpeak/**
+  - tests/modules/macs2/callpeak/**
+
 malt/build:
   - modules/malt/build/**
   - tests/modules/malt/build_test/**

tests/modules/macs2/callpeak/main.nf

@@ -0,0 +1,31 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { MACS2_CALLPEAK } from '../../../../modules/macs2/callpeak/main.nf' addParams( options: ["args": "--qval 0.1"] )
+include { MACS2_CALLPEAK as MACS2_CALLPEAK_CTRL    } from '../../../../modules/macs2/callpeak/main.nf' addParams( options: ["args": "--qval 0.1"] )
+include { MACS2_CALLPEAK as MACS2_CALLPEAK_BED     } from '../../../../modules/macs2/callpeak/main.nf' addParams( options: ["args": "--format BED --qval 1 --nomodel --extsize 200"] )
+
+workflow test_macs2_callpeak_bed {
+    input     = [ [ id:'test', single_end:false ], // meta map
+                  [ file( params.test_data['homo_sapiens']['pacbio']['genemodel1'], checkIfExists: true)],
+                  []]
+
+    MACS2_CALLPEAK_BED ( input, 4000 )
+}
+
+workflow test_macs2_callpeak {
+    input     = [ [ id:'test', single_end:false ], // meta map
+                  [ file( params.test_data['homo_sapiens']['illumina']['test_paired_end_name_sorted_bam'], checkIfExists: true) ],
+                  []]
+
+    MACS2_CALLPEAK ( input, 40000 )
+}
+
+workflow test_macs2_callpeak_ctrl {
+    input     = [ [ id:'test', single_end:false ], // meta map
+                  [ file( params.test_data['homo_sapiens']['illumina']['test_paired_end_name_sorted_bam'], checkIfExists: true) ],
+                  [ file( params.test_data['homo_sapiens']['illumina']['test2_paired_end_name_sorted_bam'], checkIfExists: true) ]]
+
+    MACS2_CALLPEAK_CTRL ( input, 40000 )
+}

tests/modules/macs2/callpeak/test.yml

@@ -0,0 +1,38 @@
+- name: macs2 callpeak test_macs2_callpeak_bed
+  command: nextflow run tests/modules/macs2/callpeak -entry test_macs2_callpeak_bed -c tests/config/nextflow.config
+  tags:
+    - macs2
+    - macs2/callpeak
+  files:
+    - path: output/macs2/test_peaks.narrowPeak
+      md5sum: d41d8cd98f00b204e9800998ecf8427e
+    - path: output/macs2/test_peaks.xls
+      md5sum: 762383e3a35e1f9ac3834fd6b2926092
+    - path: output/macs2/test_summits.bed
+      md5sum: d41d8cd98f00b204e9800998ecf8427e
+
+- name: macs2 callpeak test_macs2_callpeak
+  command: nextflow run tests/modules/macs2/callpeak -entry test_macs2_callpeak -c tests/config/nextflow.config
+  tags:
+    - macs2
+    - macs2/callpeak
+  files:
+    - path: output/macs2/test_peaks.narrowPeak
+      md5sum: 2e4da1c1704595e12aaf99cc715ad70c
+    - path: output/macs2/test_peaks.xls
+      md5sum: 5d65cb3dbd5421ea3bb5b490a100e9a4
+    - path: output/macs2/test_summits.bed
+      md5sum: 26f0f97b6c14dbca129e947a58067c82
+
+- name: macs2 callpeak test_macs2_callpeak_ctrl
+  command: nextflow run tests/modules/macs2/callpeak -entry test_macs2_callpeak_ctrl -c tests/config/nextflow.config
+  tags:
+    - macs2
+    - macs2/callpeak
+  files:
+    - path: output/macs2/test_peaks.narrowPeak
+      md5sum: 653e1108cc57ca07d0f60fc0f4fb8ba3
+    - path: output/macs2/test_peaks.xls
+      md5sum: bf86546faa7b581b5209c29b22046a0a
+    - path: output/macs2/test_summits.bed
+      md5sum: 4f3c7c53a1d730d90d1b3dd9d3197af4